| الوصف: |
The eruption of a peralkaline rhyolite magma has never been observed, yet these eruptions are amongst the most common in the Main Ethiopian Rift Valley since 1 Ma and dominate the eruption record of volcanoes that have undergone caldera collapse. The unusual rheological properties of peralkaline rhyolites, in combination with the lack of direct observations of eruptions, means that the style and hazards associated with them is essentially unknown. With 1 million people living within 10 km of a volcano in Ethiopia, and numerous geothermal power stations being built directly on these volcanoes, understanding their eruptive style and hazards is timely, and essential to robustly assess risk. This thesis aims to evaluate the eruptive styles and pyroclastic density current (PDC) hazards of post-caldera peralkaline rhyolite eruptions. I focus on Aluto volcano, a restless caldera system which has seen a multitude of peralkaline rhyolite eruptions over at least the past 16 Ka. By studying the deposits from these eruptions, I attempt to evaluate the styles of eruptive activity these magmas undergo; whether they generate PDCs, how these PDCs are generated, and how mobile they might be. I find that eruptions at Aluto occur across the edifice, akin to a monogenetic field, and that each eruption tends to undergo a very similar eruption sequence, albeit over a range of magnitudes. Eruptions typically begin with the formation of an eruption column, generating tephra fall deposits. Whilst investigating these deposits we have discovered, described and investigated a largely unrecognised type of pyroclast which I term a ‘pumiceous achnelith’. Thermal and ballistic modelling of these pumiceous achneliths indicates that pumice cones are generated by pyroclastic material falling from the sides of these columns, accumulating around the vent. Towards the end of the eruption, the eruption-column becomes unsteady, repeatedly collapsing and re-establishing, generating multiple PDCs. These PDCs usually have a high particle-concentration at their base, and tend to be confined to drainages. In most cases, the final stage of the eruption is marked by the effusion of a silicic lava flow, though it is uncertain how explosive this phase is. Using these insights, I have selected analogue PDC data, combined with modelled collapse heights and a kernel-density vent-susceptibility model, to inform a simple energy-cone model which estimates the inundation footprints of hypothetical PDCs. I employ a Monte-Carlo approach to evaluate a full range of probable eruption scenarios, and I find that the caldera, and its NW, N, and SE flanks, are particularly prone to inundation by PDCs. I combine this with geospatial data of people and infrastructure around Aluto to evaluate the collective risk, and the risk to individuals posed by PDCs. In terms of collective risk, I find that although most PDCs are constrained to within a few kilometers of the edifice, though it is still possible for more distal settlements to be inundated during rarer high-magnitude events. The population of these distal settlements is much denser than in local settlements, meaning that the collective risk is often similar between proximal and distal locations. For the individual, PDC risk is much higher closer to the edifice. I frame this risk amongst ‘everyday’ risks experienced by individuals in Ethiopia, and establish that for residents on the volcano, the time-averaged yearly risk of death by PDC is comparable to that of death by malaria, house-fire, malnutrition or road traffic accident. Using current best-practices, I have produced PDC hazard maps for different stakeholder groups at Aluto. Though there is still great uncertainty surrounding the styles and hazards associated with peralkaline rhyolites globally, this work shows that they’re capable of producing intense eruptions (intensity 7-10); generating moderate to tall eruption columns (3-16 km), which can collapse to form pyroclastic density currents with the potential to devastate local settlements and infrastructure. |
